Unable to access custom ESP32 over Serial/USB

[ESP_Sprite]

What's the problem with 'firmware erase'? Still the same issue where it can't connect? If you force GPIO0 low (e.g. solder in a bodge wire from there to GND) does it manage to erase the flash?

[rbshah7388]

Hi, shorting GPIO0 to GND also does not help. But I am getting these logs in putty now.

Code: Select all

rst:0x10 (RTCWDT_RTC_RESET),boot:0xb (HSPI_FLASH_BOOT)
flash read err, 1000
ets_main.c 371

[rbshah7388]

Finally solved it, but I will have to redesign the board! I had selected GPIO2 for I2C SCL signal that was pulled up and found that it has to be left floating or GND for ESP32 to enter to boot mode.

I only have GPIO23, 13 spare now and have GPIO15 as SDA. So hope I can use any of 23 or 13 as I2C SCL?

P.S.: Need 2 I2Cs in my design - so IO22, IO21 are already used as 1 I2C.

[sergiomarina]

Dear Friends,
I just received ESP32-WROOM-32E (rev3).
I tested two and both and I sadly found out that they do not work. Here below a selection of the output messages I got:

Code: Select all

rst:0x1 (POWERON_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:2
load:0x3fffeba4,len:4
load:0x4009f000,len:3248
entry 0x4009f574
�OHAI�ets Jul 29 2019 12:21:46
.......
rst:0x3 (SW_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0�Z;���U��=���ets@Ju����bΫX�=B��=�)�B�
@P8iH"� DB�ets Jul 29 2019 12:21:46
......
rst:0x10 (RTCWDT_RTC_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x0`,d_drv:`T�S[������ets@!��U�U����ets�ձJHUY��T:�DP����(QLNP12:21:46
....

I have never seen this before.
Replacing the power supply or the serial to USB transceiver does not change the behavior.
In steady condition there is no output on serial. Touching ESP32 pins with fingers ("EN" pin reacts the most) triggers a sequence of messages in output as described here above.

What could be the reason of both ESP32's failure?
Thanks for helping.

[ESP_Sprite]

What's your power supply (and importantly: can it deliver at least 500mA), what's your schematic?

[sergiomarina]

Hi ESP_Sprite,
thank you for your reply.
In both cases I carried out, the power supply is provided by this chain: PC's USB port + a 5v to 3.3v converter (AMS1117 3.3 LDO).
Therefore the two cases I tested are conceptually very similar they only physically change the AMS1117 3.3 LDO.
Case1 uses a separate module (there are many on marketplace) and case2 uses the 3.3 output provided by an ESP32 dev kit board.
The reason of the latter is to see if the ESP32 mounted on dev kit would have failed due to a lack of power. As a result the ESP32 on dev kit kept on working regularly while the ESP32 module under test kept on sending error messages.
The former case should increase the available current to the ESP32 under test, but the resulted error messages did not change.

I can stop using the PC's USB (which is the same power source I've used in both cases described here above) by switching to a 2A 5VDC power supply (still connecting a 5 to 3.3 converter) so we can add a test to our suite.
Do you consider valuable this additional test?

Added
I did the test as described above anyway and it did not change the result. Can we conclude that the cause is somewhere else than power supply?
In case, any idea of the failure reason?

Added
the ESP32 modules under test (I received the packet yesterday - the dealer is LCSC) have a peculiarity: the pin row at the bottom is different than usual. The label are as follows: "IO2, IO15, NC, NC, NC, NC, NC, NC, IO13, GND". "NC" labels are unusual to me.

[ESP_Sprite]

I'd still like to see your schematic if possible; especially the fact that touching EN seems to change behaviour is very suspect. From what I remember, the -E modules indeed have the flash pins not connected.

[sergiomarina]

Hi ESP_Sprite,
thank you for your reply.
Since the schematic is very simple I thought sending a picture would be slightly better (here attached)
Is that ok to you? Pls, let me know.

[ESP_Sprite]

Okay. That was not what I was expecting. If you build your device like that, I'd suggest you still follow the minimum requirements for an ESP32 module: make sure all the bootstrap pins are connected or pulled up or down correctly, and have a RC network on the EN pin to provide power-up reset. Suggest you refer to the datasheet of the module for more info. (In this case, I indeed do not see a RC network on the EN pin, and I think that explains the behaviour you see.)

[sergiomarina]

Hi ESP_Sprite,
thank you for your reply.

Pls, help me in clarifying the description reported in this doc
https://www.espressif.com/sites/default ... de__EN.pdf
pag10's picture does not seem to connect EN pin. Am I confused?

I'm aimed to setup a minimal test bench to classify ESP32 WROOM PCBs (the "minimal" requirement helps in reducing external factors)

Added
I found out the devkit pullups with a 12k resistor. I modified accordingly and now I still have RTCWDT_RTC_RESET error blinking about every second.
Regarding to the figure in the testing guide here above and the RTCWDT_RTC_RESET error I still have, is the setup in the picture I've attached missing other connections?